In order to build Grid resource sharing environment, one of the basic problems is to locate one or more entities (e.g. CPU, memory, OS, applications) that match a provided description; this is called resource discovery process. This paper proposes a model of Grid resource discovery that is based on classifying the Grid nodes according to some resource criteria. Each group has a leader (Head) that collects the information of the group members (nodes). The paper also proposes a resource discovery algorithm that optimizes resource query on the Grid network. Discussion shows how this model can satisfy the Grid resource discovery features such as: Decentralization, scalability, dynamism.